A database may be replicated on multiple nodes. These nodes may be located together or may be distributed in multiple locations. When a database is replicated on nodes that are located at a single location, a node that sits in front of other nodes may distribute requests to the nodes according to availability or other factors. This approach, however, does not scale well, may impose additional burdens for security, and may not be well suited when the nodes are distributed in multiple locations or even when nodes are located at a single location.
The subject matter claimed herein is not limited to embodiments that solve any disadvantages or that operate only in environments such as those described above. Rather, this background is only provided to illustrate one exemplary technology area where some embodiments described herein may be practiced.